Calculating the usage probabilities of statistical usage models by constraints optimization

Abstract

The systematic generation of test cases from statistical usage models has been investigated recently for specific application domains, such as wireless communications or automotive applications. For Markov chain usage models, the expected usage of a hardware/software system is represented by transitions between usage states and a usage profile, meaning probability values that are attached to the state transitions. In this paper, we explain how to calculate the profile probabilities for the Markov chain usage model from a set of linear usage constraints and by optimizing a convex polyhedron that represents the constrained solution space. Comparing the computed probability distributions of our polyhedron approach with the maximum entropy technique, which is the main technique used so far, illustrates that our results are more obvious to the intented constraint semantics. In order to demonstrate the applicability of our approach, workflow testing of a complex RIS/PACS system in the medical domain was carried through and has provided promising results.